Dynamic Impact Response

Dynamic Impact Response of Human Cadaveric Forearms Using a Wrist Brace

Richard M. Greenwald, PhD, Peter C. Janes, MD, Stephen C. Swanson, MS and Thomas R. McDonald

From the *Orthopedic Biomechanics Institute, Salt Lake City, Utah and High Country HealthCare, P.C., Vail/Frisco, Colorado

The purpose of this study was to compare the dynamic impact response of braced and unbraced cadaveric wrists using a commercially available wrist guard. Twelve arms were harvested from six cadavers. Each pair of forearms, one with and one without brace, were impacted using a modified guillotine-type drop fixture placed over a force platform. Using piece-wise linear regression analysis, we identified four phases of dynamic loading in the vertical force profile before fracture. These phases included an initial linear loading phase starting at impact, followed by a nonlinear phase, a second rapid linear loading phase, and a final nonlinear loading phase to failure. Three transition points were identified that defined the boundaries of the linear loading phases. Vertical force and impulse were significantly higher (P<0.01) at each transition point and at failure in all braced specimens compared with unbraced specimens. However, the most noticeable differences were found during the initial two loading phases. Time to each transition point and to failure was not significantly different (P > 0.27) between the braced and unbraced wrists. The results of this study differ from those obtained under more quasistatic loading conditions. Dynamic impact testing suggests that wrist guards may have a prophylactic effect during low-energy dynamic impact situations.

The popularity of snowboarding and in-line skating has increased dramatically over the last 10 years. As a result, concomitant increases in injury rates associated with these sports have also been documented. epidemiologic data have revealed that wrist injuries account for 59% of all in-line skating injuries and 19% of snowboarding injuries (Refs. 3, 10; unpublished data, data Z.J. Idzikowski and P.C. Janes, 1997). Although many participants in these sports use wrist guards, the data do not elucidate the effectiveness of wrist guards in preventing wrist fractures. Most reports in the literature encourage the use of wrist guards to prevent injury. However, recent biomechanical studies have been inconclusive regarding the efficacy of wrist guards in preventing wrist injuries.

A recent study using cadavers found no differences in failure load between pairs of wrists with and without commercially available wrist guards. Although the cadaver model may be an effective method of comparing the effects of wrist guards, the loading rate used that study (25 mm/sec) was probably substantially lower than loading rates experienced during actual fall situations. Another study documented the fracture patterns of braced and unbraced cadaver wrists during dynamic impacts. The wrist guards altered the patterns of fractures and ligament injuries, suggesting that the mechanism of injury may be a function of the loading patterns seen by wrist and the forearm. No previous studies have examined the dynamic loading curves caused by impact of the distal segment of the arm as a function of wrist guard use. The purpose of this study was to characterize the dynamic impact response of braced versus unbraced cadaver wrists.

Twelve arms sectioned below the elbow were harvested from six fresh-frozen cadavers (mean age, 47.1 years; range, 20 to 68). Radiographs were taken before testing to verify the absence of abnormalities or previous fractures. No specimens were omitted based on these criteria. The soft tissues were removed from the proximal 10 cm of each forearm, and the bones were embedded in a low melting point alloy (Affiliated Metal, Salt Lake City, Utah) for attachment to a drop fixture. A simple random number generator was used to generate a list of the forearm pairs. Using this list, we place a commercially available wrist guard design for snowboarding on either the right or left wrist of each pair of forearms. A size range of wrist guards (small to extra large) was available to ensure proper fit on each specimen. The wrist guards were constructed of Kydex (Kleerdex Co., Mount Laurel, New Jersey) and had a ventral splint from the metacarpophalangeal joints to the midforearm, while allowing range of motion of all digits. The wrist guards ere secured using three VELCRO straps (VELCRO USA Inc., Manchester, New Hampshire) positioned at the distal end of the device, the wrist joint, and the proximal end of the device (Fig. 1).

Each specimen-mounting interface complex was secured to the mounting stage n the drop fixture as shown in Figure 2. The drop fixture consisted of a gravity-driven, modified, guillotine-type track placed over an AMTI force platform (AMTI Corp., Watertown, Massachusetts). The effective mass of the drop complex was 23.0 kg. This mass was chosen because it corresponds to one-third of the body mass of an average human, representing the portion of the upper body that would be directly above the arm in a backward fall.

The mounting stage was machined such that each forearm was positioned at an angle of 75 degrees with respect to the force platform (see Fig. 2). The wrists of each specimen (both unbraced and braced) were positioned in about 40 degrees of dorsiflexion and 10 degrees of internal rotation (pronation). This particular wrist angle was produced by proper fitting of the brace with no preload and was also a likely orientation of an unbraced wrist before fall. In addition, the internal rotation angle and mounting stage were identical to those used in a previous study. To maintain this wrist position, the fingers of each specimen were secured using a nylon cord placed at the level of the interphalangeal joints. To simulate the natural landing surface for wrists during snowboarding, a foam pad with material properties very similar to packed powder snow was fixed to the surface of the force platform.

Each specimen was dropped from a height of 40 cm onto the force platform. Using the law of conservation of momentum (potential energy/mass X acceleration due to gravity X height = kinetic energy/one-half mass X velocity squared), and neglecting air resistance, the vertical velocity of the drop complex at impact was estimated to be 2.80 m/s. A height of 40 cm was determined to be a good estimate of the weight of the wrist at the beginning a backward fall. It also should be noted that although the horizontal component of a person's linear velocity during a fall situation can be quite variable, the vertical component of his of her velocity depends on entirely on the height from which the fall takes place.

The analog output from the force platform during each was sampled at 4 kHz using a 12-bit analog-to-digital converter (ComputerBoards, Inc., Middleborough, Massachusetts) interfaced to a microcomputer. The raw vertical force data from each drop were then stored on a magnetic disc for further analysis. Medial and lateral radiographs of each specimen were taken after the drop sequence. Fractures were graded according to the Frykman classification by a sports medicine fellowship-trained orthopaedic surgeon.

Initial inspection of the vertical force versus time series data revealed four phases of loading preceding failure. Two of the loading phases were linear, and the other two phases displayed nonlinear characteristics. Piece-wise linear regression analysis of the vertical force-time series data was used to identify transition points between the four loading phase. Figure 3 provides a graphical representation of the piece-wise linear regression method used in determining the three transition points for a selected trial.

The analysis protocol consisted of fining a range of data points from each linear loading phase that resulted in the best fit of the regression line (highest r2 value). It should be noted that the range of points extended over at least 95% of the linear loading phase. The r2 values ranged from 0.97% to 0.99 %. The transition points were determined by extending the range of data points used in the regression analysis in a step-wise manner to the first data point that caused a noticeable decrease in the r2 value (0.01 or more). Using such step-wise regression, point OA was determined to be the end point of the first linear loading phase, and points OB and OC marked the beginning and end of the second linear loading phase, respectively. It should also be noted that points OB and OC were determined separately by extending the range of data points of the second regression line in the direction of the phase transition.

The transition points (OA, OB and OC) and the failure point (OD) from each trial were noted. The vertical load and time from initial impact at each transition point were recorded. Impulse (intregral of the force-time curve) and the vertical velocity (calculated as impact velocity minus [impulse at point of interest/mass of drop complex] using the impulse/momentum relationship: force X change in time = mass X change in velocity) of the drop complex were calculated at each point of interest. Difference scores (braced versus unbraced) for each pair were then calculated from the recorded data. Paired t-tests were used to identify significant differences in these scores. The alpha level was set a priori at P < 0.05.

Figure 4 illustrates the vertical ground-reaction force at phase transition points OA, OB, OC, and at the failure point OD. The peak loads experienced by the forearms fitted with wrist guards were significantly higher than the corresponding loads in the group without wrist guards (3808 271 N versus 2821 763 N: P < 0.01). The duration of time from initial contact to each of the transition points was not significantly different (P>0.27) between forearm groups (Fig. 5). Figure 6 provides a comparison of the impulse applied by force platform to the entire drop complex before failure of the wrist braced versus unbraced. Impulse was significantly higher (P<0.02) in the braced condition at each of the transition points and at failure.

The fracture patterns after impact were varied (Table 1) and included fractures of the ulner shaft (3), distal radius (2), scaphoid (3), lunate (1), ulnar styloid (2), and the midradial shaft (3). Distal radial fractures were evident in four of six braced wrists compared with two of six unbraced specimens. There were three cases of fractures that included the ulnar styloid. One of the unbraced specimens had no discernible fracture.

The results of this study reveal that the use of wrist guards significantly altered the dynamic loading characteristics in cadaveric wrists, including peak load to failure and reduction in impact velocity. These results differ from the findings of a similar study using more quasistatic loading rates. We believe that dynamic testing provides a more realistic approximation of the loading rates that may be associated with falls in snowboarding and in-line skating. The combination of mass (one-third body weight) and drop height (40 cm) used in this study was intended to provide an estimate of the impact forces associated with an average forward fall in snowboarding. Although they did not measure forces during impact, Moore et al. have shown that the use of wrist guards reduces the number of fractures and ligament injuries. However, that study did not attempt to assess the ability of this brace in preventing or localizing wrist fractures. We believed it was more important to understand how the brace functions in dynamic impact situations to alter the load and the loading rate to the underlying bone and soft tissue structures. The failure loads in both the braced and unbraced conditions were substantially higher than those obtained with quasistatic loading . The loading response of viscoelastic material such as none is rate dependant, with a measurable increase stiffness with increasing loading rate. The slower loading rates used in the Giabcobetti et al. study may have been directly related to the lower failure loads recorded in both the braced and unbraced conditions. The lack of a difference in failure load between the braced and unbraced conditions in that the study may be representative of the fact that the wrists were driven to maximum dorsiflexion through a relatively slow, constant displacement rate. Thus, the ability of the brace to absorb and transmit energy was not likely a factor in the ultimate failure load. No details on the displacement (or time) to failure were reported.

In our study, the implementation of the wrist guard appeared to have the most effect in the early phases of the loading period. Point OB defines the point at which there is a sharp increase in the vertical force applied to the test system. Figure 7 provides a comparison of vertical force and time to failure for a pair of wrists. Vertical force was significantly higher from initial impact to point OB in the braced condition, suggesting that the relatively stiff brace material was resisting wrist dorsiflextion. Conversely, the time from impact to point OB was approximately the same for braced versus unbraced. As a result, the impulse (force applied over time) applied to the arm by the impact surface before point OB was significantly higher in the braced condition. The total mass of the drop fixture was basically the same between conditions (wrist guard mass was 100g for small and 130 g for large). A derivation of Newton's second law (force X change in time = mass X change in velocity) states that a higher impulse applied to the same mass will cause a greater reduction (decrease in velocity) in the momentum of a moving body.

Figure 8 Compares the vertical velocity of the drop complex and time to failure in the braced and unbraced conditions. Although the vertical velocity was lower throughout the entire loading phase in the braced condition, the patterns were similar after approximately 19.0 ms (approximately point OB). This suggests that the greatest reduction in momentum as a result of brace use occurred during the first two loading phases.

Qualitative examination of the loading characteristics suggests that the brace had little effect after point OB. Although the force and impulse variables were significantly higher in the braced condition, the loading patterns from point OB to point OC were very similar in both test conditions (that is, linearly increasing with similar slopes). This suggests a load threshold above which the brace is no longer able to decrease the momentum of test system. Based on the results of this study, it is likely that fractures that have occurred while wearing wrist guards resulted from high-energy impacts where the loads applied exceeded the capabilities of the brace. During such impacts, the brace may fail to stiffen the wrist sufficiently and prevent maximum dorsiflextion. This hypothesis concurs with the findings of Calle' and Eaton that wrist fractures occurred even when in-line skaters were wearing wrist guards. The wrist guard used in this study may provide a prophylactic effect during low-level impact situations, although it may be difficult to determine the type of fall in which the applied loads would remain below the load threshold of the brace.

The fracture locations and patterns found in this study differ slightly from results of previous dynamic impact studies, but display a similar amount of variability. These differences can be attributed to different drop conditions, limb orientation, age of the specimens, differences in brace design, and the high variability associated with fracture patterns and locations seen clinically.

There were several limitations to this study. Although the drop complex allowed identical loading that most contralateral wrists, the multiaxial loading that most likely occurs during an actual fall cannot be simulated with a device that applies only a single, planar load. Mayfield has shown that deviations in ulnar orientation or intercarpal supination result in different patterns of wrist injury. While other wrist positions could have been used , fixation in such positions would introduce more variability to the experiment. Also, the wrist orientation selected for this study was similar to that previously used in the literature and allowed better comparison of the results obtained in those studies. The cadaver wrist represent the worst-case loading scenario during a fall, but the effects of muscle tension and evasive maneuvers due to kinesthetic reaction cannot be replicated. Although a wide size range of wrist guards was available to allow proper fit on each specimen, variation in tensioning the VELCRO attachments of each wrist guard could have introduced variation into the experiment or altered the mechanical properties of the brace. Wrist kinematics during impact would also provide extremely valuable insight into the dynamic loading situation. We hypothesize that wrist fractures occur at or near terminal dorsiflexion, and that the actual load (above a certain level) may not be as important as the position of the wrist during a fall. Dynamic testing has revealed that this particular wrist guard design may be effective in preventing wrist injuries due to low-energy impacts. The addition of three-dimensional wrist kinematics alone or the addition of an inverse dynamics analysis would be extremely beneficial in future studies. The energy absorption capabilities of both the brace and the wrist joint during impact would allow a more precise assessment of the actual function of the brace. Additional testing should also be performed at subfailure loads using different drop height and effective mass combinations. Accumulation of these loading a kinematic data over a range of impact energies would allow development of a surrogate or computer model of the wrist during dynamic loading situations. Surrogate models have proven successful in parametric evaluation of knee braces. Continued development of computer models using forward dynamic simulations of the motions and loading of joints would be a more efficient means of investigating a wide variety of situations without the variability of cadaver models.

In summary, dynamic loading seems to elicit different responses in the cadaver model and may be more appropriate in simulating fall situations and in assessing the function of wrist guards in the cadaver model. Impulse and force data revealed that current wrist guard designs may have some prophylactic effect at lower impact energies because they provide additional resistance to falling motion during initial loading. However, they appeared to have little effect in reducing loading rates at higher loads. This finding provides a possible mechanism for the results of epidemiologic studies that reported wrist injuries occurred even while wearing wrist guards. Modification of fracture locations and patterns as a function of the material and mechanical properties of the bone may be possible. More detailed analysis using dynamic cadaver impact models is needed to further understand the injury mechanism and to provide a basis for improved wrist guard designs.

1. Banas MP, Dalldorf PG, Marquardt JD: Skateboard and in-line skating fractures: A report of one summer's experience. J Orthop Trauma 6: 301-305, 1992 2. Calle SC: In-line skating injuries, 1987 through 1992 (letter). Am J Pub Health 84: 675, 1994 3. Calle SC, Eaton RG: Wheels-in-line skating injuries. J Trauma 35: 946-951, 1993 4. Davidson TM, Laliotis AT: Snowboarding injuries, a four-year study with comparison with alpine ski injuries. West J Med 164: 231-237, 1996 5. Erickson AR, Yasuda K, Beynnon B, et al: An in vitro dynamic evaluation of prophylatic knee braces during lateral impact loading. Am J Sports Med 21: 26-35, 1993 6. France EP, Paulos LE, Jayaraman G et al: The biomechanics of lateral knee bracing. Part II. Impact response of the braced knee. Am J Sports Med 15: 430-438, 1987 7. Frykman G: Fracture of the distal radius including sequelae-shoulder-hand-figure syndrome-disturbance in the distal radio-ulnar joint and impairment of nerve function. A clinical and experimental study. Acta Orthop Scand (suppl) 108: 1-153, 1967 8. Giacobetti FB, Sharkey PF, Bos-Giacobetti MA, et al: Biomechanical analysis of the effectiveness of in-line skating wrist guards for preventing wrist fractures. Am J Sports Med 25: 223-225, 1997 9. Jacques LB, Grzesiak E: Personal Protective equipment use by in-line roller skaters. J Fam Pract 38: 486-488, 1994 10. Janes PC, Idzikowski J: Snowboarding injuries, in Skiing Trauma and Safety: Ninth International Symposium. ASTM STP 1182. Philadelphia, ASTM, 1993, pp 225-261 11. Mayfield JK: Mechanism of carpal injuries. Clin Orthop 149: 45-54, 1980 12. Mecham MD: Incidence and severity of head impact during free style aerial ski jumping. Master's thesis. University of Utah, Salt Lake City, Utah 1997 13. Moor MS, Popovic NA, Daniel JN et al: The effect of a wrist brace on injury patterns in experimentally produced distal radial fractures in a cadaveric model. Am J Sports Med 25: 394-401, 1997 14. NEISS Estimate Reports: Estimates for 1993-5. Washington DC, US Consumer Product Safety Commission: National injury Information Clearinghouse, 1995 15. Schieber RA, Brabche-Dorsey CM, Ryan GW: Comparison of in-line skating injuries with roller skating and skateboarding injuries. JAMA 271: 1856-1858, 1994 16. Vincent JFV: Structural Biomaterial. Princeton, NJ, Princeton University Press, 1962, pp 191-195 17. Winter DA: Biomechanics and Motor Control of Human Movement. New York, John Wiley & Sons. 1990

Peter C. Janes, MD is a provider for Vail • Summit Orthopaedics in Frisco, Edwards, and Vail, Colorado.